Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).
Inventory Control and Learning for One-Warehouse Multi-Store System with Censored Demand
Recep Bekci, Mehmet Gumus, Sentao Miao
McGill University, Canada
We study an inventory control problem called the One-Warehouse Multi-Store (OWMS) problem when the demand distribution is unknown. The OWMS system is ubiquitous in supply chain management, yet its optimal policy is notoriously difficult to calculate even under the complete demand distribution case. In this work, we consider the OWMS problem when the demand is censored, and its distribution is unknown a priori. Results show that our approach has great theoretical and empirical performances.
The privacy-preserving data-driven inventory management
Lorraine Yuan, Elena Belavina, Karan Girotra
Cornell University/Cornell Tech, United States of America
Using customer data for inventory management can improve profits and service but also increase privacy risk. We developed privacy-preserved adaptations for two data-driven newsvendor pipelines and analyzed the tradeoff between privacy loss, profits, and consumer surplus. We show that the joint approach outperforms the two-step approach. By accounting for downstream optimization problems, we can obfuscate customer data with more targeted noise injection, making it less costly in terms of profits.
Learning to order for inventory systems with lost sales and uncertain supplies
1Hong Kong University of Science and Technology; 2University of Illinois Chicago; 3New York University
Problem: Stochastic lost-sales inventory control with uncertain supply and demand is computationally challenging. We propose an efficient online learning algorithm for unknown distributions. Our algorithm achieves a regret of O(L+\sqrt{T}) when L≥log(T), outperforming existing literature. We address censored data using a coupling argument. Our method eliminates suboptimal solutions.